Jozsef Fiser

Research of Jozsef Fiser focuses on how structured visual information is acquired and converted into sophisticated internal representations for guiding cognition and behavior. He employs an integrated approach with three main components: human psychophysical and learning experiments, computational modeling of perception and learning, and multi-electrode recordings from behaving animals. His research topics include the formation of hierarchical object representations, active learning and meta-learning, teaching, complex decision making, probabilistic computation in the brain, and the relationship between beliefs and knowledge. The overarching goal of his work is to develop a statistically grounded and biologically plausible framework linking low-level visual processes and mechanisms (e.g., orientation coding and adaptation) with the development and learning of higher-level complex features and constancies, enabling efficient representations of objects and scenes for intelligent interaction with the environment.